fasrt

A blazing fast, zero-copy subtitle parser and writer for SRT and WebVTT in Rust.

Installation

[dependencies]
fasrt = "0.2"

Features

• Zero-copy, zero-allocation parsing — borrows directly from the input string
• #![no_std] support (with optional alloc and std features)
• Lazy iterator-based parsing — blocks are yielded on demand
• DFA-based lexing via logos for fast tokenization
• Strongly-typed newtypes (Hour, Minute, Second, Millisecond, Percentage) with compile-time validation
• W3C WebVTT spec conformant — validated against Web Platform Tests

SRT

• Strict and lossy parsing modes
• Timestamps (HH:MM:SS,mmm)
• Multiline cue bodies
• Writer (std feature)

WebVTT

• WEBVTT signature and header text
• Timestamps (short MM:SS.mmm and long HH:MM:SS.mmm forms, unbounded hours)
• Cue identifiers (zero-copy &str)
• Cue settings (vertical, line, position, size, align, region)
• NOTE, STYLE, REGION blocks
• Full REGION definition parsing (id, width, lines, regionanchor, viewportanchor, scroll)
• Float percentages (e.g., 50.5%)
• CRLF, CR, LF line endings
• BOM handling
• Error recovery (--> in cue body, malformed timing lines)
• Writer with round-trip fidelity (std feature)
• Cue text parsing — two-layer design:
  • CueParser: logos DFA-backed, zero-alloc token stream (no_std)
  • CueText: W3C spec-compliant DOM tree builder with Node/TagNode types (alloc/std)
  • Tags: <b>, <i>, <u>, <c>, <ruby>, <rt>, <v>, <lang>, with classes and annotations
  • W3C tree building algorithm: implied end tags, <rt> scoping, unterminated tag handling
  • Full HTML5 named character reference support (2,231 entities via phf perfect hash map)
  • Numeric ( ) and hexadecimal ( ) character references
  • Lazy text normalization via CueStr with OnceCell-cached decoding and NULL (U+0000 → U+FFFD) replacement

Optional dependencies

Feature	Default	Description
std	Yes	Enables std::io writer and thiserror::Error impls
alloc	No	Enables CueText DOM tree and entity decoding without std
memchr	Yes (via alloc/std)	SIMD-accelerated fast path for entity decoding

Benchmarks

Measured on Apple Silicon with cargo bench (Criterion).

SRT

Benchmark	Input	Time	Throughput
Parse (strict)	2 cues, 89 B	~170 ns	520 MiB/s
Parse (strict)	26 KB file	~38 µs	661 MiB/s
Parse (lossy)	332 files, ~8 MB	~12.1 ms	646 MiB/s
Collect into Vec	26 KB file	~40 µs	616 MiB/s

WebVTT

Benchmark	Input	Time	Throughput
Parse	2 cues, 96 B	~318 ns	291 MiB/s
Parse	Settings + region + style, 354 B	~915 ns	387 MiB/s
Parse	All WPT fixtures, ~34 KB	~113 µs	314 MiB/s
Collect into Vec	Settings + region + style, 354 B	~973 ns	364 MiB/s

Cue Text

Benchmark	Input	Time	Throughput
Parse	Tags only, 166 B	~316 ns	552 MiB/s
Parse	500 timestamps, ~11 KB	~14.1 µs	776 MiB/s

Run benchmarks yourself:

cargo bench

License

fasrt is under the terms of both the MIT license and the Apache License (Version 2.0).

See LICENSE-APACHE, LICENSE-MIT for details.

Copyright (c) 2026 FinDIT Studio authors.